Permalink
Commits on Nov 15, 2017
  1. add mockr no NEWS (#686)

    krlmlr committed with hadley Nov 15, 2017
Commits on Nov 13, 2017
  1. add test for capturing in non-UTF-8 locale (#680)

    krlmlr committed with hadley Nov 13, 2017
  2. Re-send revdep emails

    hadley committed Nov 13, 2017
  3. revdep results, CC @hadley

    krlmlr committed with hadley Nov 9, 2017
  4. Fix encoding issues in output recording and comparison (#679)

    gaborcsardi committed with hadley Nov 13, 2017
    * Fix expect_output and capture_output encoding
    
    We root of the problem is thar `capture_output` cannot
    capture UTF-8 output, it is just not possible currently
    with R. The output will always be in the native encoding.
    
    If UTF-8 characters are outputted, then R automatically
    converts them to the native encoding, using the equivalent
    of `enc2native`.
    
    `testthat` currently requires that the test files are
    in UTF-8, and also that in `expect_output` UTF-8
    characters are outputted. R will convert the UTF-8
    characters to native encoding, and `testthat` does
    the same now with the expected output, so they can
    be compared on all platforms.
    
    This commit:
    * Adds an `encoding` argument to `read_lines`, to
      be able to mark the re-read output as `unknown`,
      which mean the native encoding. (It was erroneously
      marked as UTF-8 before.)
    * Uses `encoding = "unknown"` in `expect_output`.
    * Adds a test case for UTF-8 characters in `expect_output`.
    
    * Fix encoding in `expect_known_output`
    
    See the previous commit about `expect_output`
    for details. `expect_known_output` still
    relies on the reference output having an UTF-8
    encoding, but also makes sure that it is
    converted to the native encoding for comparison.
    
    * Fix reporter test cases on non-UTF-8 platforms
    
    There are two issues here. The first one is that we need to
    make sure that we compare the strings in the same (native)
    encoding. So we
    * mark the actual output as in the native encoding,
      instead of (the default) UTF-8, in `read_lines`, and
    * convert the reference output to the native encoding,
      from UTF-8.
    
    The other issue is that we need to force Unicode (UTF-8)
    characters in `cli`, otherwise we can't compare the output
    with the reference output. This is because `cli` is smart
    and avoids omitting UTF-8 characters on non-UTF-8 platforms
    But the UTF-8 -> native conversion performed by R
    is not the same as the fallback mechanism of `cli`, so
    the output is different. By making sure that the `cli`
    output is UTF-8, both strings will have the same
    UTF-8 -> native conversion and they are comparable.
    
    * expect_known_output: warn for non-UTF-8 reference
    
    This can easily happen if the output is recorded on
    a non-UTF-8 platform. We also warn if the updated
    output is not UTF-8.
    
    This (with the previous commits) closes #672.
Commits on Nov 12, 2017
  1. Remove NOTE regarding is_reference() (#678)

    krlmlr committed with hadley Nov 12, 2017
    * turn off note regarding is_reference()
    
    * bump required rlang dep
Commits on Nov 2, 2017
  1. Use symbols & spinner from CLI

    hadley committed Nov 2, 2017
    Fixes #652
  2. use_tidy_description

    hadley committed Nov 2, 2017
  3. Print overall duration at end

    hadley committed Nov 2, 2017
  4. Only teardown + gc if needed

    hadley committed Nov 2, 2017
    Fixes #651
  5. Terminate early if test file doesn't exist

    hadley committed Nov 2, 2017
    Fixes #660
  6. Fix markdown styling

    hadley committed Nov 2, 2017
    Closes #653
  7. Update docs

    hadley committed Nov 2, 2017
  8. Don't promote retirement cycle on CRAN (#668)

    lionel- committed with hadley Nov 2, 2017
    Closes #662
  9. fix typo in source_file() documentation (#670)

    BarkleyBG committed with hadley Nov 2, 2017
    Fixes #669
  10. Parse in UTF-8, always, really fixes #605 (#671)

    gaborcsardi committed with hadley Nov 2, 2017
Commits on Nov 1, 2017
  1. Update test results after restyling

    hadley committed Nov 1, 2017
  2. Consistent check for encoding

    hadley committed Oct 17, 2017
Commits on Oct 25, 2017
  1. Code style (#663)

    krlmlr committed Oct 25, 2017
    * scope = "spaces", strict = FALSE
    
    * scope = "indention", strict = FALSE
    
    * manual change
    
    * scope = "line_breaks", strict = FALSE
    
    * scope = "tokens", strict = FALSE
    
    * scope = "tokens", strict = TRUE
    
    * manual style tweaks
Commits on Oct 18, 2017
  1. Handle case when `env_test$package` is `NULL` (#646)

    jimhester committed with hadley Oct 18, 2017
  2. Use cyan for times instead of white (#650)

    jimhester committed with hadley Oct 18, 2017
Commits on Oct 17, 2017
  1. avoid compiler warning (closes #649)

    kevinushey committed Oct 17, 2017
  2. Note about mockery

    hadley committed Oct 17, 2017
Commits on Oct 16, 2017
  1. revdep run, CC @hadley

    krlmlr committed Oct 16, 2017
  2. Preparing for revdep announcements

    hadley committed Oct 16, 2017
    * Remove old code
    * Define email template data
  3. Use local labelling instead of rlang equiv

    hadley committed Oct 16, 2017
    Includes some code formatting improvements
  4. Updates after review of changes

    hadley committed Oct 16, 2017
  5. Force first argument in quasi_label

    hadley committed Oct 16, 2017
    In order to generate improved error stacks
  6. Correct link slugs in README

    hadley committed Oct 16, 2017
  7. Merge branch 'f-revdep'

    hadley committed Oct 16, 2017
  8. add revdep results, CC @hadley

    krlmlr committed Oct 16, 2017
Commits on Oct 13, 2017
  1. add simple tests

    kevinushey committed Oct 13, 2017